Interview with Arundhati Roy by Amy Goodman and Nermeen Shaikh, Democracy Now

Writer Arundhati Roy discussed her new essay, "Apocalypse Now and Then," published in Equator magazine. Roy delivered remarks on artificial intelligence (AI) earlier that week at the People's AI Summit in New York. In the interview, she explained the essay's title and argued that AI concentrates control over language, labor, natural resources, and human cognition. She also outlined the actions she considers necessary.

The Title: Warnings as Marketing

Roy said the essay began with the public statements of Jacob Coxon, a whistleblower who left Anthropic. Coxon warned that the company was building something resembling "the Manhattan Project on steroids." Roy said the scientists involved were driven by profit and by the goal of reaching the technology first.

She stated that AI executives presented the future of the human species as dependent on their decisions, while assuring the public that they would act responsibly. Roy acknowledged that AI carries real risks. She noted that executives have put the chance of AI eliminating the human race at around 10 percent.

Roy argued that the apocalyptic framing also served a commercial purpose. In her view, the more lethal and irreducible the technology appeared, the more valuable the companies became, particularly ahead of planned initial public offerings. She described the approach as a sales pitch: "Apocalypse Now, by us, get yours free."

Publishing, Authorship, and Consent

Roy said AI has disrupted the publishing industry. She described publishers examining the work of new writers with the scrutiny of detectives inspecting a crime scene. She cited a series of scandals in which literary prizes went to works later found to have been produced by AI. She also reported that a new prize for AI-written work had been introduced.

Roy said her own work is among those of hundreds of thousands of writers that Anthropic used, without permission, to train its large language models. She said that after a class action lawsuit, she received notice that writers would be offered approximately 2,000 dollars per book. She rejected the compensation as inadequate for work that, in her words, represents a lifetime of developing a language, a culture, and a way of thinking.

Roy also said AI companies have taken people's voices, likenesses, and paintings in the same manner. She criticized the ease with which users can request text "in the style of" a specific author.

She cited Sam Altman, chief executive of OpenAI, who said: "We see intelligence as a commodity to be sold by the meter." Roy argued that the collective intelligence of humanity has been appropriated and commercialized by "a handful of white men." She said mathematicians face the same problem that writers face.

Language, Trust, and Social Cohesion

In the essay, Roy wrote that disrupting the relationship between humans and language would undermine social cohesion. She cited the Gospel of John, which begins, "In the beginning was the word, and the word was with God, and the word was God." She said she cited the text as a writer, not as a statement of religious belief.

Roy said that language includes spoken communication and storytelling, and that the most widely read religious texts, including the Bible and the Quran, are built from stories. She said AI converts this material into products.

Roy described a loss of trust in authorship. She said she has seen videos of herself saying things she has never said. She warned that people can no longer assume that a text or statement is genuine.

She raised concerns for younger generations. She asked whether they would have the opportunity to strive, fail, produce imperfect work, or experience the difficulty and satisfaction of creating art, literature, or music. Roy said she sees a "tidal wave of mass psychosis" forming. She compared the experience to looking in a mirror and seeing a likeness that does not respond as oneself, with no reliable way to tell whether the image or the person is real.

Roy said the central harm is not limited to the destruction of the human race. She said the larger danger is the destruction of humanness.

Capitalism and Control

The essay carries the subtitle "Artificial Intelligence is Capitalism's Final Destination." The interviewers asked whether the central issue is the technology itself or who controls it. Roy said both matter. She said the technology is significant, but she framed AI as the latest stage of a long process of corporate expansion.

Roy said she has followed corporatization and neoliberalism for years, through the cases of mining and dam construction. She described large dams as a method of taking control of a river, centralizing its water, and deciding who receives it. She said mining extracts water and minerals, and that uranium extraction destroys mountains without regard for the ecosystem effects. She said AI extends this extraction to "the mind and the organs of human beings."

Roy compared consumer AI applications to the shells left on a beach when the sea receded before a tsunami. She listed the roles the applications claim to fill, including therapist, mother, lover, novelist, and translator. She said these offers conceal a larger threat.

Roy said that the people who control AI development have openly discussed overthrowing governments and financial systems. She quoted one figure in the essay who said, "Everyone on the internet needs to be policed and stack ranked."

She predicted mass unemployment. She described workers in India who wear cameras on their foreheads to record their movements. The recordings train robots that will replace them. Roy said these workers cannot refuse, because millions of people are waiting to take their jobs.

Roy said five or six individuals who control the technology have, in her words, disaffiliated from the human race. She said they have built homes in remote areas stocked with food, water, and ammunition, and that they are discussing data centers in space.

Roy said AI would drain human vitality, spirit, and soul. She called it "actual idiocy."

Reversibility and Regulation

The interviewers said that the outcome may not be inevitable, but that AI is unlikely to be reversed. They asked how the industry could be regulated and what response is required.

Roy disagreed that AI is either irreversible or inevitable. She said the industry is currently facing a structural crisis. Data centers and supercomputers require large amounts of land, money, data, energy, and water. She said these requirements make the companies unviable at their current scale.

Roy said the companies are attempting to tie themselves to governments. She said they are positioning themselves as essential to national security. She said they have offered to produce weapons and lab-grown viruses, and to protect governments from unspecified threats.

Roy said that the money invested in AI has been distributed so widely that a collapse would cause a broader financial crisis. She said the companies are trying to become "too big to fail," but that they will fail. She said the central question is how society will absorb the consequences.

The interviewers compared the situation to the dot-com bubble, in which most companies failed while a few, such as Google, survived. Roy said she does not know whether all AI companies will fail or how the situation will develop. She said the portrayal of AI as a miraculous engine of progress is false.

Data Centers in India

The interviewers asked about the environmental effects of data centers, including their water and electricity demands, and about their location in India.

Roy said data centers are being built in Visakhapatnam. She said authorities have granted clearances as if the facilities were ordinary buildings, without assessing their effects on surrounding areas or on groundwater. She said the project faces significant local opposition.

Roy said that companies could avoid resistance by building in sparsely populated regions such as Ladakh. She warned that such construction would threaten the glaciers, which she described as the world's water tower.

Roy said people in the United States are resisting data centers. She said companies may relocate facilities to countries such as India that are willing to accept the environmental costs, as has happened in the past with toxic industrial waste.

Roy said everyone is on the front line of this development, not only communities near data centers. She said people usually recognize such threats too late.

Recommended Action

Roy said people should spread awareness of the threat. She said they should be willing to refuse. She rejected the view that AI is already too powerful to challenge. She said the public must recognize the problem and state it clearly.

Roy said she would accept the consequences of stating her position. She said that whether people are helpless is a choice, and that the first step is to name the problem publicly.

Excerpt from "Apocalypse Now and Then"

Roy read the closing section of the essay during the interview:

While there's so much debate about what AI can and cannot do for humans, there is too little talk about what it's doing to humans.

Studies have warned of a relationship between AI and cognitive decline in human beings. Human intelligence looks set to plunge in proportion with the rise of artificial intelligence. Perhaps our unused brains will atrophy, our heads will shrink, and our fingers will turn into undulating prompt tentacles.

Those of us who've already lived our lives in a different kind of world know what we're losing.

But what of the generation that is about to come of age? Will it have its own language, its own thoughts, its own grief? Will it have the right to strive, to fall, to fail, to be imperfect? Will it experience the impossible pain and ecstatic pleasure of solving a math equation, or creating art, or literature, or music of its own, however flawed?

The triumph of AI will be a crime against humanness. It'll bottle our brains in formaldehyde. It'll destroy the world as we know it. The men who promote it and profit from it are cognizant of this and have already made arrangements to exit into another realm. The oligarchs are announcing the battle lines proudly, loudly, that it's us against them. They suggest without flinching that for AI to thrive, humanity may have to die.

The young generation that is rising up to fight it understands this.

I know that earlier in this essay, I suggested that we should, quote, regulate with extreme prejudice. But what I really want to say about those AI companies is this, terminate with extreme prejudice.